Like The Hub?
Join our community.
Join

Dylan Marando: Hope in Health Care: Embracing AI could save our health systems

Commentary

This article is the first installment in a four-part series, “Hope in Health Care“, exploring how we can modernize and adapt our health systems for the future.

Where have you gone, Harold Johns?

For some (including Simon & Garfunkel), it’s Joe DiMaggio. For others, it might be Jean Beliveau, Julie Andrews, or Rosa Parks. Whether you lived in the era or not, you likely regard the mid-twentieth century with some romanticism—and its cultural figures with heroic esteem. The age of moonshots and microwaved dinners was a moment of transition and growth, while also being a time of boundless optimism. The world moved in giant steps. And almost everyone appeared convinced that, if given the chance, they could keep up.

In one corner of our country, an enterprising physicist by the name of Harold E Johns was convinced that Canadians could beat cancer. And then he proved himself right.

While the Tommy Douglas types were mapping out the future of health-care administration, big brains such as Johns were overhauling medical science itself. More specifically, Johns and colleagues at the University of Saskatchewan were doing pioneering work in the emerging field of radiation therapy. They channeled a frenzy of activity in nuclear research into the first ever accurately calibrated dose of a life-saving treatment. In 1951, a Saskatchewanian mother of four with cervical cancer became the first patient in the world to benefit from this breakthrough. She lived another 47 years.

In his harnessing of a novel science, Johns turned technological uncertainty into something truly wonderful. He stood on the precipice and saw the mountain peaks ahead. He did something that brought hope and healing to an estimated 70 million patients across the globe. And in the decades that followed, countless successors have built on the legacy of Johns and others by forging further breakthroughs in radiation therapy, advancing the field far beyond its position in 1951. 

And, thus, in our own moment of rich technological potential, with a mix of nostalgia and confidence, our health-care leaders might ask: Where have you gone, Harold Johns? Or, perhaps more appropriately, what should we be doing to support the “Johns-ians” among us today? What would Harold Johns do amid lively debates on artificial intelligence in health care, for example? What support would he need to bring his ideas from the laboratory into patient care?

On the issue of AI in health care, I’d like to think that Johns would be bullish on the convergence of new technologies and public policy. I think Johns would find common cause with present Canadian luminaries David Naylor and Geoffrey Hinton (yes, that Geoffrey Hinton) who suggested in 2018 that appropriately managed deep learning had the potential to transform health-care delivery (although Hinton has since added cautionary notes to his earlier championing). I think Johns, backed by a lifelong commitment to higher education, would encourage our generation of researchers to do the hard work of seeking and scaling the next mountain range, yet again making Canadian science and technology a marvel of the medical world.

Of course, one could (and should) argue that this call to action is more than a mere thought experiment. It’s, for many, a present reality. At the University of Toronto, where Johns founded the Department of Medical Biophysics following his time in Saskatchewan, there are already researchers setting the pace for AI in health care, testing interventions such as the use of patient voice and machine learning to diagnose disease. At the Princess Margaret Cancer Centre, another organization shaped by Johns’ legacy, researchers are finding ways to leverage data and digital technologies to enhance triage and better match patients with clinical trials. More broadly, AI is proving to be a valuable tool for improved detection of breast cancer as well as enhanced risk assessment for pancreatic cancer, among various other promising interventions. 

In other words, Harold Johns hasn’t gone anywhere. His pioneering spirit remains central to Canadian health research and technology.

But, is an ethos of transformation also evident as we migrate from science and technology to public policy? What happens when one moves from academic journals to venues of care for patients? Are we getting the full benefit of technological progress in our day-to-day experiences of health care? Are health-care policymakers responding to the complexities of AI with creativity or constraint? If the latter, which policymakers are willing and able to roll up their sleeves and break through the (ana)log jam? After all, what would Johns have done if then Saskatchewan Premier Tommy Douglas didn’t do the tedious work of prioritizing cancer care; or if the Leslie Frost government in Ontario hadn’t established the Ontario Cancer Institute? What if the vigor of technologists wasn’t met with the thoughtful planning of our political leaders?

Folks smarter than I have observed that data readiness, valid concerns around inequities, regulatory frameworks, and human resources—among other factors—can slow the progress of AI in health care across Canada. Indeed, predictions of AI-type technologies significantly reducing health human resource pressures are now decades old, yet parts of our health system are lacking in their propensity to adopt latest and greatest technologies—contributing to stories of long wait times and burnt-out health-care providers. So, it’s fair (and understandable) to say that our systems have room for improvement. And it’s also fair to say that fundamental governance changes in the midst of a technological inflection point are difficult to navigate. 

And therein lies the rub. Health-care policy systems that aren’t already purpose-built for transformation—that have yet to make a habit of developing, assessing, adopting, and scaling technological innovation, as well as its various supplements (e.g. high-quality data, and agile training programs)—are at risk of not fully capturing the enormous potential that is AI in health care. Our current policy ecosystems may become disoriented—to the point of static—if mega-opportunities, including but not limited to AI, are not systematically calibrated with our rightly complex health-care institutions.

In the case of AI in health care, some important first policy steps toward such alignment could include: credentialing that results in better training for health-care professionals in the use and management of emerging technologies; acceleration of electronic health implementation in a way that makes data voluminous, higher-quality, and more usable across providers; greater resourcing and scaling of localized problem-solving; and creating increased capacity within administrative apparatuses for true value-based procurement—inclusive of more rapid assessment and adoption of leading-edge technologies. If we can capitalize on these reforms (as starting points), we will continue to foster an environment where health-care entrepreneurs can thrive, and where patients can get the best care in the world.

So, for every ounce of legitimate public policy reactiveness and administrative complexity we inject into our health-care systems across Canada, we should inject equal parts public policy foresight. We should view an “AI in Cancer Care Moonshot” not as a possible science-fiction title, but as a practical necessity. As often as we query provincial comparisons on wait times, we should ask which province will lead the country, or even the international community, in defining the future of medical technology. The impulse toward this kind of virtuous competition is discernable in initiatives like Ontario’s Innovation Pathway, or Alberta’s new Chair in AI in Health, or Nova Scotia’s Health Innovation Hub. But more is needed. Fast.

For the past three years, many of us have felt a degree of anxiety about our health-care systems. We’ve tried to avoid getting sick. We’ve strained to hold up frayed processes of diagnosis and treatment. And we’ve grown exhausted sorting facts from fiction in popular medical discourses. 

Yet, there is another path. A more positive and aspiring path. A path that is both newer, and older. It’s the Harold Johns path.

Sean Speer: Why won’t the Liberals own their big-spending record?

Commentary

One of the oddest traits of the Trudeau government is that even after seven and a half years in office, it still cannot decide whether to boast about its fiscal profligacy or downplay it. On one hand, its extraordinary willingness to “invest” public dollars in different causes and projects is clearly core to its political identity. On the other hand, it seems self-conscious about the true costs of its progressive predisposition to public spending. 

This inherent tension is not new for the government. It goes back to its origins. The Trudeau government was first elected with a plan for deficit spending that it went some lengths to emphasize was “modest”, “prudent” and “short-term.” Even as mounting evidence contradicted these claims (as former Prime Minister Stephen Harper famously predicted), it seemed reluctant to fully embrace its own fiscal policy choices. The government and its supporters clung for some time to claims about the federal debt-to-GDP ratio and other fiscal anchors before eventually abandoning them altogether. 

Today they’re still stuck between two competing ideas: the government rhetorically rejects the so-called “austerity” of the Conservatives but it also counterintuitively argues that its fiscal policy broadly conforms to its predecessor’s. 

That leaves it to us to try to reconcile these incongruous interpretations of the Trudeau government’s fiscal policy. Let me therefore defend the government from its progressive critics and its own preposterous spin about “fiscal restraint” as Finance Minister Chrystia Freeland put it in her 2023 budget speech. 

There’s nothing restrained about the government’s fiscal policy unless the word has assumed such an elastic definition as to render it essentially useless as a description of government budgeting. Notwithstanding the claims of its critics and defenders, the Trudeau government is easily the biggest-spending government in modern Canadian history. 

The COVID-pandemic obviously complicates an analysis of the government’s fiscal record. The unprecedented spike in federal spending—it increased by nearly 80 percent between 2019-20 and 2020-21—needs to be accounted for. Yet even if one opposed the magnitude of overall pandemic spending or specific spending choices, it’s fair to say that any government would have increased program spending and run budgetary deficits of some size in response to the pandemic. 

Fiscal figures beyond 2020-21 are also still only projections that the government may overshoot. It has a long track record of pushing up spending and the size of the deficit as the budget’s outer-year projections become closer to the present. The 2023 budget for instance revised the government’s previous projection that it would be back in surplus in 2027-28. The Parliamentary Budget Office’s analysis now suggests that it may happen by 2035. The point here is that there’s good reason to assume that current spending projections for 2023-24 and beyond are bound to be adjusted upward in the coming years.     

Notwithstanding these limitations, we can still look backwards and forwards to evaluate the Trudeau government’s fiscal policy and the tension that’s run through how it has spoken about spending, deficits, and debt.  

Let’s start with program spending prior to the pandemic. It grew by an average of 6.2 percent per year over the Trudeau government’s first four years in office. This stands in contrast with the Harper government’s last four years which actually saw program spending decline by an annual average of 0.9 percent (see Figure 1). 

Graphic credit: Janice Nelson.

Then there’s the pandemic response which drove up federal spending to unprecedented levels. Just consider: the 2020-21 deficit itself was more than 10 percent larger than the entire federal budget in the last year of the Harper government. Another way to think about it is: its predecessor could have collected no revenues in its final year in office and still ran a smaller deficit than the one recorded in 2020-21. 

Emergency spending as a share of GDP reached 18.5 percent that year which though it matched the G-7 average (see Figure 2) was among the highest in the world including peer jurisdictions such as Australia, New Zealand, and the United Kingdom. 

Graphic credit: Janice Nelson.

The government’s post-pandemic projections envision program spending falling relative to a pandemic-induced high but it’s notable that it doesn’t fully revert to its pre-pandemic trajectory. After a nearly 80-percent increase in 2020-21, the government projects a roughly 25-percent reduction over the next two years before it resumes growing again. A considerable share of the pandemic spending has simply become part of the government’s ongoing expenditure baseline. 

As I’ve outlined in a past Hub article, if the pandemic had never happened and the government simply kept growing spending at roughly 6 percent per year as it had prior to the pandemic, program spending in the current fiscal year would be as much as $45 billion lower than is currently projected (see Figure 3). Such back-of-the-envelope analysis provides a sense of how much the pandemic spike has altered the course of the government’s own pre-pandemic trajectory. 

Graphic credit: Janice Nelson.

As mentioned earlier, there’s good reason to be skeptical about the post-2023-24 projections. They anticipate a level of spending restraint that the government has never delivered. They’ve also already changed a great deal in the past 16 months alone.

Consider for instance that between the 2022 Budget, the 2022 Fall Economic Statement, and the 2023 Budget, projected program spending in 2024-25 alone has gone up by nearly $25 billion (see Table 1). Another way to put it is: between the past two budgets alone, the government’s own projection for program spending in 2026-27 was pushed up three years to 2024-25. Even if one is prepared to grant the government some dispensation due to pandemic uncertainty, it’s difficult to defend such significant movement—particularly in the later years—over such a short period. 

Sources: Graphic credit: Janice Nelson.

These figures suggest that, as these outer years get closer, it’s more likely than not that the government’s own spending projections will grow. The interplay between outstanding policy priorities, political exigencies, and ideological preferences undoubtedly points in the direction of higher spending and more deficits and debt in the coming years. 

The Trudeau government’s spendthrift assumptions were evident at its beginning and they remain evident today. At this point, the only ones unable to see it are those blinded by some combination of ideology or partisanship. The government might as well accept it. The rest of us have.